Resource · Glossary · AI data center operations

    What Is CMDB Integration?

    CMDB integration is the set of connections that keep a configuration management database accurate and useful: discovery tools and other systems feed it configuration items and relationships, and ITSM, monitoring and reporting tools read from it. Enrichment fills in attributes from the best source; northbound interfaces pass records up to the systems that act on them.

    A CMDB on its own is a list. Integration is what keeps the list true and puts it to work. For the basics of the database itself, see what a CMDB is.

    Why integrate

    Why a CMDB needs integration

    A CMDB kept up to date by hand goes stale as soon as hardware moves, a card is swapped or a job lands on a different node. The value of a CMDB comes from two directions: accurate data coming in, and other systems using it going out.

    In an AI data center the configuration items go beyond servers and switches. Accelerators, clusters, resource pools, jobs, containers, tenants and projects all need to be related, so that a fault on one GPU can be traced to the job, the tenant and the power feed involved.

    Directions

    Inbound, enrichment and northbound

    • Inbound discovery: hardware read through BMCs over Redfish and IPMI, network devices over SNMP, virtualization and cloud platforms through their APIs, and workload schedulers such as Slurm and Kubernetes for jobs.
    • Enrichment: attributes discovery cannot see, or that another source knows better: rack and U position, warranty dates, owner and cost center, tenant and project.
    • Northbound: records and events passed up to the systems that act on them: ITSM, alerting, reporting and finance.
    CMDB INTEGRATIONNorthbound: systems that use the recordsITSMincidents,changesAlertingimpact,webhookReportscost,auditCMDBconfiguration items, relationships, ownersDiscoveryBMC, SNMP,RedfishSchedulerjobs,tenantsCloudsAWS, Azure,GCP, VMwareSources: discovery and enrichment
    Discovery and enrichment sources feed the CMDB from below; ITSM, alerting and reports use its records from above.
    Northbound interface

    The northbound interface

    The term comes from network management. A platform's southbound interface faces the devices it manages; its northbound interface faces higher-level systems. In CMDB terms the northbound side carries records and events upward: APIs that let another tool query configuration items and relationships, webhooks that push an event to a URL the receiving system owns, and ITSM connectors that link incidents and changes to the affected items.

    The same word is used for the reverse case. Northbound collection means a platform reads other platforms through their open APIs, for example pulling resources, capacity, performance and alerts from cloud platforms into one view.

    ITSM

    CMDB and ITSM integration

    Linking ITSM to the CMDB is where integration pays back. An incident tied to a configuration item inherits its relationships, so impact analysis can list the services, jobs and tenants affected. A change request can be checked against dependencies before it is approved. Repeated failures on one hardware model show up in problem management. Correlating alerts before they reach ITSM matters too: one fault should open one ticket, not one ticket per alert.

    Existing CMDB

    Integrating with a CMDB you already run

    Many organizations already have a CMDB of record, often ServiceNow. Adding a hardware-level source then becomes an enrichment job: decide which system is authoritative for each class and attribute, identify items by a stable key such as serial number or UUID, and let the target's reconciliation (in ServiceNow, the Identification and Reconciliation Engine) match incoming items so they are not duplicated. Avoid two-way overwrites of the same field, and agree how often data syncs.

    ServiceNow is a trademark of its respective owners. Sensaka is not affiliated with, endorsed by, or sponsored by that company. Comparisons reflect publicly available product information and common deployment patterns at the time of writing; features, packaging and pricing change, so verify current capabilities against each vendor's own documentation before making a purchasing decision.

    In Sensaka

    CMDB integration in Sensaka

    The Sensaka AI Infrastructure CMDB discovers devices and components through BMCs and network protocols, records rack and U position, and tracks relationships between workloads, containers, nodes and accelerators, together with tenant, project and owner. Changes are kept as historical snapshots, and impact analysis follows the relationships. Sensaka Sync loads discovered inventory into the Sensaka CMDB and compares it with your existing asset records.

    On the northbound side, alerts can be routed to PagerDuty, OpsGenie, Slack, Jira or a webhook, incidents and changes run through Sensaka's ITSM service management, and metrics from AWS, Azure, Google Cloud, OpenStack and VMware come into the same view. If you already run a CMDB of record, the interfaces for your project are agreed during scoping. Everything runs on-premises.

    FAQ

    Common questions

    What is CMDB enrichment?

    Adding attributes to configuration items from the source that knows them best: serial numbers and firmware from the server's BMC, rack and U position from the data center model, warranty dates from the vendor, owner, tenant and project from the scheduler or the service catalog. Reconciliation rules decide which source wins for each attribute.

    What is the difference between northbound and southbound integration?

    Southbound integration reads the devices below a platform: servers through their BMCs, switches over SNMP, power and cooling systems. Northbound integration connects the platform to the systems above it, such as ITSM tools, alerting, reporting and other management platforms.

    How does a CMDB integrate with ServiceNow?

    Usually through ServiceNow's import and integration tooling, with its Identification and Reconciliation Engine matching incoming configuration items to existing records so they are not duplicated. The decisions that matter are which system is authoritative for each attribute, how a CI is identified (serial number, asset tag or UUID) and how often data syncs.

    Why does an AI data center need CMDB integration?

    Because its questions cross systems. Which jobs and tenants does a failed accelerator affect? Which project used a rack's GPU-hours? Answering them needs the hardware, the scheduler and the ticketing tool to agree on the same configuration items and relationships.

    Try it on your own hardware

    The Sensaka pilot is free for 30 days and runs on-premises, inside your data center, with success criteria agreed before it starts. Integrators and resellers can also ask about the partner programme, which pays a 25% margin in the first year.